Spectral analysis of the flow signals of the space-periodic capacitance Transducer

This work investigates the application of the Burg maximum entropy method (MEM) to the problem of measuring the velocity of solids flowing in pneumatic conveyors. The technique is based on the use of a space-periodic transducer (SPT) which generates flow signals with a spectral content dependent on the velocity of the flowing particles. An earlier technique utilized the periodogram method to estimate the spectra of the SPT signals. The performance of the periodogram method and MEM is first compared by utilizing sinusoidal waveforms corrupted by additive white Gaussian noise. The results show that the MEM produces better resolution with accurate frequency and power estimates even for short data records. The MEM is then employed to estimate the power spectral densities of actual flow signals consisting of six records each having 256 data points sampled at a rate of 2000 sps. Three different criteria for selecting the order of the prediction error filter are adopted. It was found that the MEM spectra contain a highly prominent main lobe with a dominant spectral component which leads to measured velocities between 2-4 m/s. The distortion in the main lobe and the minor peaks which marred the periodogram spectra of these records disappear in the MEM spectra for properly-selected orders. Application of MEM to flow signals 'reveals that it is a promising
method for solid flow measurement.